Introduction

Aptamers are single-stranded DNA (ssDNA) or RNA molecules containing tens of nucleotides that can bind certain targets with high specificity and affinity (Famulok et al. 2000). Since 1990, many aptamers for various targets from small molecules to whole cells have been selected and widely used in designing biosensors, biomedicine and other fields (Famulok and Mayer 2014; Wang et al. 2016; Toh et al. 2015). The aptamers were selected using a process called Systematic Evolution of Ligands by Exponential Enrichment (SELEX) (Famulok et al. 2000; Ellington and Szostak 1990; Tuerk and Gold 1990). Some major steps were involved in this process: incubation of the random nucleic acid library with the target, separation of the bound library elements from the unbound ones, amplification and purification (Toh et al. 2015; Blind and Blank 2015). Among them, the most important step for a successful selection is to separate the bound sequences from a large number of unbound sequences. For small-molecule targets, the tiny differences of size, weight and charge properties between target-binding complexes and unbound sequences complicate the separation step. In addition, different from targets such as biomacromolecules and cells, the aptamers selected for small-molecule targets generally show relative lower affinities based on the published data of aptamers (McKeague et al. 2015).

At present, most of SELEX methods for small-molecule targets typically involve the immobilization step (McKeague and DeRosa 2012), such as conventional SELEX and capture SELEX as the two representative methods. The conventional SELEX needs the immobilization of target molecule (Ellington and Szostak 1990), which may alter the original conformation of the target molecule, and cause difficulty or even failure to the selection due to the steric hindrance of the interface. Moreover, this strategy is not suitable for small molecule without active groups for immobilization. The capture SELEX process is immobilizing the library instead of the target molecule by designing a defined docking sequence in the random region of the library, which can be hybridized to the capture oligonucleotide on the magnetic beads (Stoltenburg et al. 2012). However, if the potential aptamer is free in solution in subsequent applications, the tethered library may cause variations in the shape and conformation of the potential aptamer, thus affect the binding of the aptamer (Ruscito and DeRosa 2016). The emergence of SELEX methods based on nanomaterials without the immobilization step may provide a better choice for small-molecule targets. These methods are performed in solution without the need of target immobilization, in which the ssDNAs adsorbed on nanomaterials by non-covalent interactions such as hydrogen bonding, π-π binding or electrostatic interaction will be dissociated from the nanomaterials by target-binding, for example, the most used graphene oxide (GO)-SELEX (WooáKim and BockáGu 2012; Nguyen et al. 2014; Gu et al. 2016).

Gold nanoparticles (AuNPs) exhibit many similar properties as GO for DNA adsorption (Liu 2012). The mechanism of ssDNA adsorption on AuNPs has been well studied, including van der Waals force between oligonucleotides and AuNPs and strong electrostatic interaction caused by dipolar interaction, which depends on the configuration and orientation of ssDNA (Li and Rothberg 2004a, b). From now on, many colorimetric aptasensors based on AuNPs have been developed due to their simplicity and high sensitivity (Li and Rothberg 2004a; Yang et al. 2011; Peng et al. 2013). In the presence of targets, the aptamers are folded by target-binding and are desorbed from the surface of AuNPs, resulting in the aggregation of AuNPs in salt solution, and thus the color of the solution changes from red to purple-blue (Liu et al. 2014). However, to the best of our knowledge, the selection of aptamers using AuNPs has not been reported.

In this work, assuming that AuNPs can help to separate the bound sequences from the unbound sequences in a random library, we attempt to select DNA aptamers for small-molecule targets based on AuNPs using the porphyrin compound Zinc(II)-Protoporphyrin IX (ZnPPIX) as the target. ZnPPIX is a trace amount of normal metabolite formed in heme biosynthesis with a molecular weight of only 626.051 g/mol. The final reaction of biosynthetic pathway is the chelation of iron with protoporphyrin. During iron deficiency or compromised iron utilization, zinc becomes an alternative metal, resulting in increased formation of ZnPPIX (Labbé et al. 1999). As shown in Fig. 1, the chemical structure of ZnPPIX is similar with Fe(III)-Protoporphyrin IX (hemin) and N-methyl mesoporphyrin IX (NMM), and the aptamers of ZnPPIX may also interact with hemin and NMM. The SELEX progress can be monitored by the recovery rate and fluorescence enhancement of NMM after each round selection. And we further examined the interactions of the obtained aptamer with NMM and hemin to confirm its potential as a light-up fluorescent probe or a DNAzyme. The feasibility of AuNP-based SELEX strategy in selecting aptamers for small-molecule targets will be confirmed by this work.

Fig. 1
figure 1

Structures of porphyrins used in this study. a Zinc(II)-Protoporphyrin IX (ZnPPIX). b N-methyl mesoporphyrin IX (NMM). c Fe(III)-Protoporphyrin IX (Hemin)

Full size image

Experimental Section

Chemicals

ZnPPIX and NMM were purchased from Frontier Scientific Inc. (USA). Hemin was purchased from Sigma-Aldrich (China). 10 × PCR buffer, deoxynucleotide triphosphates (dNTPs), rTaq polymerase, 20 bp DNA Marker, 6 × loading buffer, and pMD18-T Vector Cloning Kit were purchased from TaKaRa Biotechnology (China). Regular agarose was purchased from Biowest (Spain). Streptavidin-agarose beads were purchased from Thermo Fisher Scientific (USA). Molecular Probes SYBR® Green I was purchased from Invitrogen Technology (China). All experiments were done with sterile Milli-Q water (18.2 MΩ cm).

Oligonucleotides

The random ssDNA library with a random region of 40 nucleotides contains two primer-binding regions (5′-ATACCAGCTTATTCAATT-40 N-AGATAGTAAGTGCAATCT-3′). The 5′ biotinylated reverse primer was used to product ssDNA (bio-p2: 5′-Bio-AGATTGCACTTACTATCT-3′). The random library and the primer sequences were all purchased from Sangon Biotech Co., Ltd. (China). The aptamers were synthesized by GENEWIZ Inc. (China).

Gold Nanoparticle-Based SELEX

13 nm AuNPs were synthesized by the citrate reduction of HAuCl4 according to the literature reports (Turkevich et al. 1951; Frens 1973).

The random library (200 pmol) was heated at 95 °C for 5 min and quickly cooled on ice for 10 min, which was conducive to the formation of secondary structures of ssDNAs. Then, it was mixed with ZnPPIX with a molar ratio of 1:100 in SELEX buffer (20 mM HEPES, pH 7.4, 60 mM NaCl) and incubated for 1 h at room temperature (RT) in the dark. One milliliter of 16.7 nM AuNPs were added to the above solution in SELEX buffer to a final volume 2 mL and further incubated for 1 h at RT. Then, the solution was centrifuged at 8000 rpm for 1 h and the supernatant was concentrated and purified by centrifugal filter from Merck Millipore Ltd. (USA). The amount of ssDNA was measured using a BioPhotometer from Eppendorf (Germany), and the recovery rate was calculated by the ratio of the amount of ssDNA in the enriched pool to the added amount of ssDNA.

The supernatant was subjected to PCR amplification by the biotinylated reverse primer. The condition of PCR was 3 min at 95 °C, followed by 6–14 cycles of 30 s at 94 °C, 30 s at 48 °C, and 15 s at 72 °C, and a final extension step for 5 min at 72 °C. The PCR product was incubated with streptavidin-agarose beads for 30 min and washed with PBS for three times to remove unbound DNA. After denaturation in 0.1 M NaOH solution for 2 min, the desired ssDNA strands were isolated from streptavidin-agarose beads. The pH was neutralized to 7.0–7.5 with 1.5 M HCl solution, followed by the purification by ethanol precipitation. The collected ssDNA pool was used as a secondary library for the next round of selection. The process is repeated until the affinity of the selected pool with the target no longer increases.

Enrichment Monitoring

The ssDNA pool of a given round was dissolved in SELEX buffer, then heated on 95 °C for 5 min and cooled down to RT. Then, 1 μM of ssDNA was mixed with 1 μM of NMM in 100 μL SELEX buffer and incubated for 1 h at RT before measurement. The excitation wavelength was set as 399 nm and the fluorescence emission spectrum from 550 nm to 750 nm was recorded on a spectrofluorometer (Hitachi, Japan).

Cloning and Sequencing

After confirming the enrichment, the chosen ssDNA pool was amplified with unmodified primers, then the PCR product was cloned into Escherichia coli by PMD18-T TA cloning kit from TaKaRa (China). A total of 35 clones were randomly selected and sent to Sangon Biotech Co., Ltd. (China) for sequencing. DNA secondary structures were further predicted by M-fold (http://mfold.rna.albany.edu/) (Zuker 2003).

Dissociation Constant (Kd) Determination

ZnPPIX (0.5 μM) was mixed with different concentrations of aptamer sequences in SELEX buffer and incubated for 1 h at RT. After incubation, the fluorescence intensity was recorded using a spectrofluorometer from Hitachi (Japan). The dissociation constant (Kd) curve fitting was performed according to fluorescence enhancement of ZnPPIX at 590 nm and DNA concentration by nonlinear fitting using Origin v8.0. The fitting formula is Y = 1 + 0.5 (Qmax−1) (Kd/C + 1 + nX/C−[(Kd/C + 1+nX/C)2−4nX/C]1/2), where Y is the ratio of the fluorescence intensities of porphyrin in the presence and absence of aptamer (F/F0), X represents the concentration of aptamer, C represents the concentration of porphyrin, and n represents the putative number of binding sites for porphyrin on aptamer. Kd and Qmax were obtained using the Origin 8.0 software (Wang and Rando 1995; Yang et al. 2019).

Peroxidase Activity Measurement

ssDNA was dissolved in SELEX buffer, denatured at 95 °C for 5 min, and then cooled down to RT. Hemin (1 μM) was mixed with 2 μM of ssDNA in SELEX buffer and incubated for 1 h at RT. Then, 2 mM H2O2 and 5 mM ABTS were added to initiate the reaction. The initial rate (V0) was calculated from the first 10 s of the reaction time curve (Δε = 36,000 M−1 cm−1).

Circular Dichroism Analysis

The obtained aptamer was dissolved in HEPES buffer (pH 7.4, 20 mM HEPES) and SELEX buffer, respectively. The samples were heated at 95 °C for 5 min and then cooled down to RT. Then ZnPPIX (3 μM) was mixed with 3 μM of aptamer in SELEX buffer and incubated for 1 h at RT. The circular dichroism (CD) spectra between 220 and 320 nm were recorded on a CD chiroptical spectrometer from Applied Photophysics (UK).

Results and Discussion

AuNP-based SELEX

To develop a novel immobilization-free platform for the selection of aptamers for small-molecule targets, we performed an overall AuNP-based SELEX process, as shown in Fig. 2. Briefly, the random ssDNA library was incubated with high concentration of targets in SELEX buffer, then the mixture was incubated with AuNPs in SELEX buffer in the presence of 60 mM NaCl to completely adsorb free ssDNA sequences, and the target-binding sequences, which were not absorbed on the surface of AuNPs, were separated from unbound sequences by centrifugation. Then, the target-binding sequences were concentrated and used as templates for PCR amplification to purify into a secondary ssDNA pool for the next round of selection. The process is repeated until the affinity of the selected pool with the target no longer increases. Finally, the detailed sequences can be determined by cloning and sequencing of the enriched pool.

Fig. 2
figure 2

Schematic illustration of the AuNP-based SELEX process

Full size image

To separate target-binding aptamer sequences with unbound ssDNA sequences and reduce non-specific adsorption, 13 nm AuNPs were synthesized by the classic citrate reduction method. The synthesis of citrate-stabilized AuNPs was determined by UV–Vis spectroscopy and TEM. As shown in Fig. 3, the as-prepared AuNPs show a maximum SPR absorption peak at 519 nm, and the uniform particle size and morphology with the average size of about 13 nm. To completely remove the non-specific non-adsorbed sequences in the ssDNA library, the adsorption conditions of the ssDNA library on AuNPs were optimized first before starting the SELEX. Considering the adsorption of DNA on AuNPs will be affected by the salt concentration in buffer and incubation time (Liu 2012), and the adsorption is accompanied by fluorescence quenching of the FAM-labeled ssDNA (Zhang et al. 2012), the fluorescence intensity of FAM-labeled ssDNA library was measured in the presence of various NaCl concentrations and for different incubation times after incubation with AuNPs. As shown in Fig. S1, the adsorption increased with the increasing of NaCl concentration and incubation time, and under the optimal adsorption conditions (60 mM NaCl and incubation time of 75 min), the FAM-labeled ssDNA library was almost 100% adsorbed on AuNPs. Thus, in the SELEX procedure, the mixture of the library and target was incubated with AuNPs in SELEX buffer in the presence of 60 mM NaCl for 75 min before centrifugation to minimize the non-specific non-adsorbed ssDNA.

Fig. 3
figure 3

Characterization of gold nanoparticles. a The UV–Vis spectrum of gold nanoparticles. b The TEM image and the particle size distribution of gold nanoparticles

Full size image

The recovery rate of the selected ssDNA pool after each round was determined to assess the enrichment efficiency of each round to determine if further selection is needed. The selected pool was concentrated before PCR amplification, and its amount was determined, then recovery rate was calculated by the ratio of the amount of ssDNA in the enriched pool to the added amount of ssDNA. The stronger the binding of the selected pool to target, the relatively more amount of ssDNA would be collected, correspondingly making a higher recovery rate. As shown in Fig. 4a, it was found that the ratio gradually increased with the increased number of the selected rounds. Until the 11th round, the recovery rate reached a maximum, and there’s no raise after one more round of selection, indicating the successful enrichment of pool. The structure of ZnPPIX is very similar with NMM (Fig. 1), so the aptamers of ZnPPIX may also interact with NMM. Besides, after binding with its specific ssDNA sequence, the fluorescence of NMM will increase (Yang et al. 2019). Thus, the fluorescence enhancement of the enriched pool after each round for NMM was also investigated to further determine the enrichment efficiency. As shown in Fig. 4b, it was found that fluorescence intensity of NMM gradually increased before 11 rounds, and the fluorescence of 12th round was slightly lower than that of the 11th round, which further confirmed the successful enrichment of the 11th pool.

Fig. 4
figure 4

Enrichment process. a Recovery ratio of the enriched ssDNA pool. b Fluorescence intensity of NMM upon binding with the enriched ssDNA pool. Error bars represent the standard deviation from three separate trials

Full size image

We randomly selected 35 clones of the 11th pool for sequencing. Finally, 34 sequences were obtained, and the detailed sequence information was summarized in Fig. 5a. Although no obvious conservative motif was identified, we observed that most of the random sequence regions are rich in guanine, which is statistically higher than the predicted 25% of the random bases, indicating that these sequences tend to form G-quadruplex structures, consistent with the previously reported aptamers for porphyrin (Li et al. 1996; Yang et al. 2019). The secondary structures of these sequences were further predicted by M-fold (http://mfold.rna.albany.edu/). Six representative structures were chosen as aptamer candidates for further analysis, named as ZnP1, ZnP6, ZnP7, ZnP13/16, ZnP18 and ZnP23, respectively (Fig. 5b). The interaction of these candidates with ZnPPIX was investigated by fluorescent spectroscopy, as shown in Fig. 5c, and ZnP1 showed the most significant fluorescence enhancement for ZnPPIX compared with the initial ssDNA library, confirming the binding ability of ZnP1 to ZnPPIX. Therefore, ZnP1 was considered to be the aptamer of ZnPPIX for further analysis.

Fig. 5
figure 5

Characterization of aptamers. a A total of 34 sequences were obtained after cloning and sequencing. No.13 is same as No.16. The sequences with red letters were chosen as aptamer candidates for further analysis. b Predicted secondary structures of the aptamer candidates by M-fold (http://mfold.rna.albany.edu/). c The fluorescence enhancement of the six chosen aptamer candidates for ZnPPIX. The concentration of ssDNA was 2 μM, and the concentration of ZnPPIX was 2 μM

Full size image

Dissociation Constant (Kd) Measurements

To be more conducive to applications in the future, the full-length aptamer ZnP1 was shortened to a 41-mer ZnP1.2 by removing the two end primers. The binding properties of ZnP1 and ZnP1.2 were further examined by continuous variation experiments, which were performed by recording the fluorescence enhancement of ZnPPIX at 590 nm under different ratios between ZnP1.2 and ZnPPIX. Similarly, for NMM, the operation was performed by recording the fluorescence enhancement at 614 nm. As shown in Fig. S2, according to the Job–Plot curves, the binding ratio between ZnP1 and ZnPPIX is 1:1 (Fig. S2a) and that of ZnP1 and NMM is 3:2 (Fig. S2b), while the binding ratio between ZnP1.2 and ZnPPIX is 1:1 (Fig. S2c) and that of ZnP1.2 and NMM is 3:2 (Fig. S2d).

The dissociation constants of ZnP1 and ZnP1.2 with ZnPPIX and NMM were further determined by fluorescent spectroscopy. As shown in Fig. 6, the apparent Kd values of ZnP1 with ZnPPIX and NMM were determined to be 13.65 ± 3.99 μM and 7.55 ± 1.24 μM, respectively, and ZnP1 exhibited a 17-fold fluorescence enhancement for ZnPPIX at the saturation concentration, while a 31-fold enhancement for NMM. The apparent Kd values of ZnP1.2 with ZnPPIX and NMM were determined to be 9.53 ± 1.86 μM and 1.35 ± 0.17 μM, respectively, and ZnP1.2 exhibited a 12-fold fluorescence enhancement for ZnPPIX at the saturation concentration, while a 28-fold enhancement for NMM.

Fig. 6
figure 6

Binding curves of ZnP1 and ZnP1.2. The dissociation constant curve of ZnP1 with ZnPPIX (a) and NMM (b). The dissociation constant curve of ZnP1.2 with ZnPPIX (c) and NMM (d)

Full size image

Evaluation of Peroxidase Activity

Some interest has been raised from previously reported porphyrin aptamers as DNAzymes with peroxidase activity (Hollenstein 2015; Travascio et al. 1998). A wide range of related applications have also been reported, such as detection of telomerase activity (Pavlov et al. 2004), screening of G-quadruplex ligands (Kong et al. 2008), and development of DNA sensors (Liu et al. 2018). We further investigated the peroxidase activity of the obtained aptamers using the 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS)-H2O2 system. The changes in absorbance at 414 nm over time can reflect the oxidation rate of ABTS. As shown in Fig. 7a, hemin alone had low-native peroxidase activity, while the ssDNA library also did not show significantly enhanced peroxidase activity. Both ZnP1 and ZnP1.2 exhibited enhanced peroxidase activity, of which ZnP1.2 was 22 times higher than that of hemin alone (Fig. 7b). As shown in Fig. 7c, the color changes of the reaction systems were also very evident after 10 min, further confirming the enhanced peroxidase activity of the obtained aptamers. We also compared the enzymatic activity of the obtained aptamer ZnP1.2 with the other reported aptamers (PS2.M and Nm2.1). PS2.M is the aptamer of hemin selected by Li et al. (Li et al. 1996) and has been used as a DNAzyme because of its excellent peroxidase activity (Travascio et al. 1998; Li et al. 2007), and Nm2.1 is the aptamer of NMM selected by our research group recently, which exhibits 19-fold higher catalytic efficiency than hemin alone (Yang et al. 2019). Both PS2.M and Nm2.1 are parallel G-quadruplexes. Before the enzymatic activity measurement, all the ssDNAs were dissolved in SELEX buffer, denatured at 95 °C for 5 min, and cooled down to RT. As shown in Fig. S3, under this condition, ZnP1.2 exhibited a slightly higher peroxidase activity than that of PS2.M and Nm2.1.

Fig. 7
figure 7

Peroxidase activity and conformational analysis. a Catalytic kinetics of hemin alone and DNA-hemin complexes. b The initial reaction rate V0 was calculated from the changes during the first 10 s (∆ε = 36,000 M−1 cm−1). Error bars represent the standard deviation from three separate trials. c Photographs of the oxidation of ABTS by H2O2 taken after reaction at RT for 10 min. d CD spectra of ZnP1.2 in different solutions

Full size image

Conformational Analysis

Based on the above results, this obtained aptamer ZnP1.2 not only recognizes its target ZnPPIX but also recognizes the other porphyrin molecules (NMM and hemin), which can be considered to be a porphyrin aptamer. The conformational analysis of ZnP1.2 was further performed by circular dichroism (CD) spectroscopy. As shown in Fig. 7d, ZnP1.2 displayed a positive peak at around 275 nm and a negative peak at around 245 nm in HEPES buffer in the absence of Na+, while the positive peak exhibited a blue shift and moved to 265 nm in the presence of 60 mM Na+ (SELEX buffer), indicating the formation of a parallel G-quadruplex structure, which is characterized of a maximum peak at around 265 nm and a minimum peak at around 240 nm (Bugaut and Balasubramanian 2008). Furthermore, the CD spectrum of ZnP1.2 in SELEX buffer could maintain characteristic peaks of parallel G-quadruplex after the titration of ZnPPIX, indicating that the interaction between ZnP1.2 and ZnPPIX is caused by the formation of G-quadruplex.

Conclusion

A novel immobilization-free SELEX method for small-molecule targets was developed by utilizing the ssDNA absorption property of gold nanoparticles. ZnPPIX was used as the target, and the progress of SELEX was monitored by the recovery rate and NMM fluorescence enhancement of the enriched pool after each round. After 11 rounds of selection, an aptamer ZnP1 and its truncated aptamer ZnP1.2 with low-micromolar dissociation constants were obtained. Moreover, both of them showed the enhancement of fluorescence for NMM and the enhanced peroxidase activity for hemin. Therefore, the functional aptamer has potential to be a light-up fluorescent probe and a DNAzyme. The structural conformation indicated the formation of a parallel G-quadruplex structure. In summary, these results suggest this simple and convenient AuNP-based SELEX can be used to select aptamers for small-molecule targets.